Analysis of the algorithm: From rankings to backup genes.

Ranking section

For each of the affected genes present in the list, a ranking with the score and percentile of every possible gene was obtained. Once this was done, the position for every known backup gene was obtained and the statistical descriptor of all the values calculated.

Table 1. Backup position values for every kernel and net before integration.

Net Kernel Elements Min Max Average Standard_Deviation
biological_process ka 70 0.0 0.698 0.075 0.158
biological_process rf 70 0.0 0.723 0.079 0.159
cellular_component ka 71 0.0 0.886 0.128 0.198
cellular_component rf 71 0.0 0.87 0.125 0.195
disease ka 24 0.0 0.946 0.137 0.26
disease rf 24 0.001 0.866 0.129 0.238
genetic_interaction_weighted ka 73 0.025 1.0 0.782 0.396
genetic_interaction_weighted rf 73 0.0 1.0 0.329 0.322
molecular_function ka 7 0.0 0.777 0.156 0.275
molecular_function rf 7 0.0 0.75 0.162 0.274
pathway ka 30 0.001 1.0 0.308 0.454
pathway node2vec 30 0.0 0.685 0.111 0.197
pathway rf 30 0.001 0.176 0.033 0.049
phenotype ka 23 0.0 0.965 0.225 0.305
phenotype rf 23 0.0 0.865 0.196 0.276
protein_interaction_unweighted ka 73 0.001 1.0 0.347 0.471
protein_interaction_unweighted node2vec 73 0.001 0.94 0.067 0.172
protein_interaction_unweighted rf 73 0.0 0.74 0.037 0.124

Table 2. Backup position values for every kernel after integration

Integration Kernel Elements Min Max Average Standard_Deviation
integration_mean_by_presence ct 1676 0.0 0.978 0.162 0.272
integration_mean_by_presence el 1676 0.0 0.944 0.412 0.262
integration_mean_by_presence ka 1676 0.0 1.0 0.069 0.154
integration_mean_by_presence rf 1676 0.0 0.914 0.055 0.12
mean ct 1676 0.0 0.978 0.132 0.244
mean el 1676 0.0 0.965 0.215 0.251
mean ka 1676 0.0 1.0 0.041 0.125
mean rf 1676 0.0 0.919 0.046 0.108

Backup Pairs Coverage By Layer

CDF plots of non integrated kernels

CDF plots of integrated kernels

NO DATA